In the ever-evolving saga of man versus machine, a group of authors has filed a class-action lawsuit against OpenAI, claiming that ChatGPT has developed the unfortunate habit of remembering their books without buying them dinner first. The suit, filed in a San Francisco federal court, alleges that OpenAI gobbled up thousands of books to train its models and then had the audacity to more or less reproduce their contents, which, the writers argue, is slightly more alarming than your average book club recap.
The plaintiffs, a literary collection of names including Paul Tremblay and Mona Awad, have politely suggested that ChatGPT outputs content which is eerily similar to their work and does so, rather inconveniently, without permission or royalties. In what can only be described as the world’s most literate court complaint, the authors accuse OpenAI of mass copyright infringement by feeding its AI models with data sets full of pirated books, an approach that certainly puts a new spin on the idea of “reading for inspiration.”
OpenAI, whose AI chatbot has a penchant for confidently incorrect answers and inventing citations from time to time, has not publicly addressed the complaint, possibly because it is still figuring out how to say “I learned it from you, Dad” in legalese.
The case highlights a murky area of copyright law in the generative AI space, where the definition of “learning” is slowly blurring the lines between AI training and literary looting. Legal experts are watching closely, presumably from behind towering stacks of legal textbooks, to see whether training AI on copyrighted work without explicit consent amounts to fair use or fair game. So far, the courts have yet to issue a definitive answer, much like ChatGPT when asked who really won the War of 1812.
Whether this lawsuit ends in a settlement, a precedent, or a muttered “my bad” remains to be seen, but for now the literary community is watching with the same trepidation one might reserve for a rogue librarian who quietly mutters plot spoilers under their breath.
And so, once again, the plot thickens, only this time the AI already read the ending.